JAX, M.D. A framework for differentiable physics*

We introduce JAX MD, a software package for performing differentiable physics simulations with a focus on molecular dynamics. JAX MD includes a number of physics simulation environments, as well as interaction potentials and neural networks that can be integrated into these environments without writing any additional code. Since the simulations themselves are differentiable functions, entire trajectories can be differentiated to perform meta-optimization. These features are built on primitive operations, such as spatial partitioning, that allow simulations to scale to hundreds-of-thousands of particles on a single GPU. These primitives are flexible enough that they can be used to scale up workloads outside of molecular dynamics. We present several examples that highlight the features of JAX MD including: integration of graph neural networks into traditional simulations, meta-optimization through minimization of particle packings, and a multi-agent flocking simulation. JAX MD is available at https://www.github.com/google/jax-md.

Spatial Analysis of Regional and Income Inequality in the United States

Understanding the spatial or geographical dependence of income inequality and regional inequality is crucial in the study of inequality. This paper employs a multi-scale, multi-mechanism framework to map and analyze historical patterns of regional and income inequality in the United States (US) by using state and regional panel data spanning over a century. To explore the patterns systematically and see the role of spatial partitioning, we organize the data around several established geographical partitions before conducting various geographical information system (GIS) analyses and statistical techniques. We also investigate the spatial dependence of income inequality and regional inequality. We find that spatial autocorrelation exists for both types of inequality in the US. However, the magnitude of spatial dependence for regional inequality is declining whereas it is volatile for income inequality over time. While income inequality has been at its peak in the most recent decades, we also notice that regional inequality is at its lowest point. As for the choice of partitioning, we observe that within inequality dominates for Census Divisions and Bureau of Economic Analysis (BEA) regions. Conversely, we see that between inequality overall contributes the most to the inequality among Census Regions.

Nuclear compartmentalization of TERT mRNA and TUG1 lncRNA is driven by intron retention

The spatial partitioning of the transcriptome in the cell is an important form of gene-expression regulation. Here, we address how intron retention influences the spatio-temporal dynamics of transcripts from two clinically relevant genes: TERT (Telomerase Reverse Transcriptase) pre-mRNA and TUG1 (Taurine-Upregulated Gene 1) lncRNA. Single molecule RNA FISH reveals that nuclear TERT transcripts uniformly and robustly retain specific introns. Our data suggest that the splicing of TERT retained introns occurs during mitosis. In contrast, TUG1 has a bimodal distribution of fully spliced cytoplasmic and intron-retained nuclear transcripts. We further test the functionality of intron-retention events using RNA-targeting thiomorpholino antisense oligonucleotides to block intron excision. We show that intron retention is the driving force for the nuclear compartmentalization of these RNAs. For both RNAs, altering this splicing-driven subcellular distribution has significant effects on cell viability. Together, these findings show that stable retention of specific introns can orchestrate spatial compartmentalization of these RNAs within the cell. This process reveals that modulating RNA localization via targeted intron retention can be utilized for RNA-based therapies.

Overset Grid Assembler and Flow Solver with Adaptive Spatial Load Balancing

An overset mesh approach is useful for unsteady flow problems which involve components moving relative to each other. Since the generation of a single mesh around all components is prone to mesh stretching due to the relative motion of bodies, using the overset grid methodology, an individual mesh can be generated for each component. In this study, a parallel overset grid assembler was developed to establish connectivity across component meshes. Connectivity information was transferred to the developed parallel flow solver. The assembler uses multiple methods such as alternating digital tree and stencil walking to reduce the time spent on domain connectivity. Both the assembler and solver were partitioned spatially so that overlapping mesh blocks reside in the same partitions. Spatial partitioning was performed using a 3D space partitioning structure, namely octree, to which mesh blocks are registered. The octree was refined adaptively until bins of octree could be evenly distributed to processors. The assembler and solver were tested on a generic helicopter configuration in terms of load balance, scalability, and memory usage.

8. The international law of the sea

The international law of the sea is one of the oldest disciplines of public international law. In fact, the identification and application of principles for governing the roughly 70 per cent of the earth’s surface that consists of water has been a topic of interest for centuries. This chapter deals with the main principles and rules that make up the international legal regulation of the seas. It begins by discussing the most important legal sources in the law of the sea, including the 1982 United Nations Law of the Sea Convention. It then discusses the spatial partitioning of the sea and the different maritime zones that exist in the law of the sea; discusses piracy; and examines a number of selected issues relating to the conservation of marine life. The final section provides a short introduction to dispute settlement in the law of the sea.

Area-Wide Elimination of Subterranean Termite Colonies Using a Novaluron Bait

We investigated the use of termite baiting, a proven system of targeted colony elimination, in an overall area-wide control strategy against subterranean termites. At two field sites, we used microsatellite markers to estimate the total number of Reticulitermes colonies, their spatial partitioning, and breeding structure. Termite pressure was recorded for two years before and after the introduction of Trelona® (active ingredient novaluron) to a large area of one of the sites. Roughly 70% of the colonies in the treatment site that were present at the time of baiting were not found in the site within two months after the introduction of novaluron. Feeding activity of the remaining colonies subsequently ceased over time and new invading colonies were unable to establish within this site. Our study provides novel field data on the efficacy of novaluron in colony elimination of Reticulitermes flavipes, as well as evidence that an area-wide baiting program is feasible to maintain a termite-free area within its native range.

Population Dynamics of Chewing Lice (Phthiraptera) Infesting Birds (Aves)

In the past 25 years, studies on interactions between chewing lice and their bird hosts have increased notably. This body of work reveals that sampling of live avian hosts, collection of the lice, and the aggregated distributions of louse infestations pose challenges for assessing louse populations. The number of lice on a bird varies among host taxa, often with host size and social system. Host preening behavior limits louse abundance, depending on bill shape. The small communities of lice (typically one–four species) that live on individual birds show species-specific patterns of abundance, with consistently common and rare species, and lower year-to-year population variability than other groups of insects. Most species of lice appear to breed continuously on their hosts, with seasonal patterns of abundance sometimes related to host reproduction and molting. Competition may have led to spatial partitioning of the host by louse species, but seldom contributes to current patterns of abundance.